By Michael White | June 9th 2008 08:15 PM | Print | E-mail
One of the most useful things about having our brains is the ability to anticipate predictable events: we can see that it's going to rain, or that it's getting dark, and prepare accordingly. Some things in life are completely random, some are almost perfectly predictable (like the sun rising tomorrow at 5:35 AM in the Midwest US), and most other things are not quite so regular as the sunrise, but predictable nonetheless. We use neural networks in our brain to anticipate these events, but anticipation is obviously not limited to organisms that have brains. How do other organisms, like the single-celled E. coli anticipate change? One way to make reliable predictions about change is to notice that one event correlates with another. Pavlov's dogs knew that the sound of the bell meant dinner was on its way. Bacteria that you ingest with your food move from an outdoor environment into the much warmer, low-oxygen environment of your gut. To survive, and just as importantly, compete against all the other bacteria in your gut, the bacterium has to quickly change its set of active genes so that it can metabolize nutrients in this low-oxygen environment. A bacterium that can anticipate this key switch to a low-oxygen lifestyle, by picking up on prior clues, will have an advantage. A research group led by Saeed Tavazoie, at Princeton, guessed that the common human gut bacterium E. coli might anticipate the impending intestinal low-oxygen conditions by taking note of the immediate increase in temperature experienced when it enters your mouth. As described in a recent paper in Science, they grew batches of E. coli in a fermentor, and looked to see whether E. coli switched on their low-oxygen genes when the scientists turned up the temperature. Just as the researchers guessed, cranking up the temperature cause the bacteria to switch on their low-oxygen genes. The bacteria anticipate an upcoming change in their environment, and they do this without a brain, just with a network of biochemical reactions. If these biochemical reactions can act like a bacterium's brain, can E. coli be taught to unlearn this increase temperature-decrease oxygen response? An individual bacterium can't, but perhaps these biochemical regulatory networks are flexible enough to evolve a new response with some Pavlovian training. Tavazoie's group subjected their bacteria to a set of conditions: instead of an increase in temperature, followed by a decrease in oxygen (the human gut scenario), they increased the temperature and increased the oxygen. A wild E. coli strain should be at a disadvantage here, since typically a temperature increase is the cue to switch to a low-oxygen mode. In their experiments, the original bacterial strain didn't do so well under these new conditions, but after a few thousand generations of evolution in a lab fermentor, the scientists were able to isolate a bacterial strain that learned the lesson: when the temperature was increased, these bacteria now anticipated the increase in oxygen and switched on the right set of genes. These bacteria grew much better under these conditions than the original strain. The big lesson here is that even in one-celled bacteria without nervous systems, biochemical regulatory networks can be tweaked by evolution to note correlations in the environment, and learn the right response. And when the right response changes, evolutionary pressure can tweak a cell's biochemistry, over a few thousand generations, to learn a new response.